PROBLEMY EKOROZWOJU – PROBLEMS OF SUSTAINABLE DEVELOPMENT

2016, vol. 12, no 1, 125-134

Rebuilding the Pillars of Sustainable Society Index:

a Multivariate Post Hoc I-distance Approach

Kształtowanie wskaźnika Zrównoważonego

Społeczeństwa: analiza wielowariantowa I-distance

Dejan Savić, Veljko Jeremić, Nataša Petrović

Faculty of Organizational Sciences, University of Belgrade, Belgrade, Serbia

E-mail (Corresponding Author): savicd@fon.bg.ac.rs

Abstract
Sustainable Society Index (SSI) is a composite indicator constructed to measure and describe societal progress
along all three dimensions of sustainable development: human, environmental and economic. In this paper, we
explore possibilities to evaluate and enhance SSI ranking calculation methodology based on the use of an iterative
multivariate post hoc I-distance approach. Based on the assessment on how each indicator contributes to the final
position of different countries and identification of the most influential indicators, we examine possibilities of
reduction of a number of indicators. The goal is to improve the stability of the ranking results and overall quality
of the model, focusing on the analysis of the relative contribution of the indicators in an iterative assessment
process. By this, we provide in-depth analysis and more comprehensive understanding of specific factors that
determines one country ranking position. Thus proposed approach can support policymakers to identify key indi-
cators and focus the priority areas where investment in improvement measures and programs would have the most
efficient impact on the overall positioning of the country.

Key words: sustainable development; Sustainable Society Index; composite indicators; I-distance; relative con-
tributions

Streszczenie
Indeks Zrównoważonego Społeczeństwa (Sustainable Society Index, SSI) jest zagregowanym wskaźnikiem stwo-
rzonym w celu pomiaru i opisu postępu społecznego w trzech wymiarach rozwoju zrównoważonego: publicznym,
środowiskowym i ekonomicznym. W artykule przeanalizowano możliwości oszacowania i rozszerzenia metodo-
logii SSI w oparciu o wielokrotną analizę wielowariantową I-distance. W oparciu o ocenę, jak każdy wskaźnik
przyczynia się do ostatecznego wyniku osiąganego przez różne kraje i identyfikację wskaźników o największym
znaczeniu, zbadamy możliwość zmniejszenia ilości branych pod uwagę wskaźników. Celem jest poprawa wiary-
godności otrzymywanych wyników i ogólnej jakości modelu, z podkreśleniem znaczenia analizy względnych
udziałów wskaźników w wielokrotnym procesie oceny. Umożliwi to przeprowadzenie szczegółowych badań i
bardziej wszechstronne podejście do poszczególnych czynników które wpływają na miejsce, które dany kraj zaj-
muje w rankingu. Proponowane narzędzie może pomóc decydentom w zidentyfikowaniu kluczowych wskaźników
i wskazać obszary priorytetowe, w ramach których inwestycje i programy modernizacyjne będą miały największy
wpływ na pozycjonowanie danego kraju.

Słowa kluczowe: rozwój zrównoważony wskaźnik Społeczeństwa Zrównoważonego, wskaźniki zagregowane,

analiza wielowariantowa, względny udział

1. Introduction losophy and economics from many centuries ago

(Mebratu, 1998). However, the most common defi-
The roots of sustainability concept could be found in nition of sustainable development was published in
different sources from the sciences of religion, phi- The Brundtland Report (WCED, 1987), where sus-

Electroniccopy
Electronic copy available
available at:
at: https://ssrn.com/abstract=2856356
https://ssrn.com/abstract=2856356
126 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134

tainable development was defined as development depth analysis of the differences between countries.
that satisfies the needs of the existing, without jeop- SSI methodology has been refined several times to
ardizing possibilities of the future generations to sat- improve conceptual coherence and statistical sound-
isfy their needs. Sustainable development includes ness, with the support of Joint Research Centre
various environmental, economic and social factors, (JRC) (Saisana and Philippas, 2012). In latest fifth
which are stated as the three pillars of the edition, the new SSI-2014 framework has been pre-
sustainability which do not exclude each other, but sented, consisting of 21 indicators, grouped into
even can accent each other (UN, 2005). Rising seven categories and three dimensions (Table 1).
interest for sustainable development is a result of the
perception of current worrying conditions of the Table 1. The SSI-2014 conceptual framework
global human environment. It calls and demands Dimen- Code
Category Indicator
sion
urgent reaction of all members of society, focused on
1. Sufficient Food var1
long-term environment protection and overall Basic
sustainable development of the humanity (Petrovic, 2. Sufficient to Drink var2
Needs
2012). 3. Safe Sanitation var3
Sustainable development is a critical and popular 4. Education var4
concept that is open to different approaches and Human
Health 5. Healthy Life var5
interpretations. Naturally a large number of Wellbe-
researchers, organizations, institutions and ing 6. Gender Equality var6
international agencies developed and offered many Personal 7. Income Distribution var7
various methodologies and concepts for measuring & Social
8. Population Growth var8
sustainability. Conceptual approaches to measuring Devel-
sustainable development could be grouped into two opment 9. Good Governance var9
general categories: (1) set of indicators and (2) com- var10
10. Biodiversity
posite indicators. Unlike the sets of indicators, Natural
11. Renewable Water Re- var11
composite indicators have the ability to summarize Re-
sources
complex and sometimes elusive processes into a sources
Environ- 12. Consumption var12
single figure to benchmark a country’s performance mental
in policy consumption (Giovannini et al., 2008). 13. Energy Use var13
Wellbe-
Numerous sustainability composite indicators were ing Climate 14. Energy Savings var14
developed and proposed in recent decades. The &
15. Greenhouse Gases var15
following approaches are mostly present and Energy
discussed: Index for Sustainable Economic Welfare 16. Renewable Energy var16
ISEW (Daly and Cobb, 1989), Genuine Progress 17. Organic Farming var17
Econ-
Indicator GPI (Cobb et al., 1995), Genuine
Econo- omy 18. Genuine Savings var18
Savings/Adjusted Net Savings GS/ANS (Hamilton
et al., 1997), Environmental Sustainability Index ESI mic var19
19. Gross Domestic Product
Wellbe-
(Esty et al, 2005), Environmental Performance Index ing
Transi-
20. Employment var20
EPI (Esty et al, 2006), Ecological Footprint EF tion
21. Public Debt var21
(Ewing et al., 2010), Human Development Index
HDI (UNDP, 2014) and Sustainable Society Index
SSI (van de Kerk and Manuel, 2014). Among all of Because of lack of a scientific basis for the
them, only SSI approach is tending to consider all attribution of different weights to the indicators,
three dimensions of sustainable development every indicator has received the same weight for the
integrally. aggregation into dimensions (SSI, 2014). In the last
The Sustainable Society Index (SSI) has been edition, dimensions are not aggregated into a single
developed since 2006 with the aim to be a value for the overall composite, in view of the
comprehensive quantitative method for negative correlation between Human and
measurement of the health of coupled human- Environmental Wellbeing (SSI, 2014). Composite
environmental systems and to describe societal indicators like SSI are designed to measure a multi-
progress across human, environmental and criteria performance, where is important to avoid
economic dimensions (Saisana and Philippas, 2012). composite indicators to represent unstable
SSI tends to overcome the main noted shortcomings assessments, as their stability ensures the amount of
of previously developed metrics: a limited definition validity of the observed system (Dobrota et al.,
of sustainability, a deficit of transparency and an 2015b). The usefulness of multivariate analysis for
absence of regular updates (van de Kerk and Manuel, the aggregation of individual indicators to a compo-
2008). SSI is calculated for 151 countries accounting site indicator is often emphasized (Saltelli et al.,
for 99% of the world population, with regular two- 2008). As weights are essentially value judgments
year updates that demonstrate developments over (Giovannini et al., 2008), results and values of
time and available underlying data that allow in- composite indicators are significantly influenced by

Electroniccopy
Electronic copy available
available at:
at: https://ssrn.com/abstract=2856356
https://ssrn.com/abstract=2856356
 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
the indicator weights and often are the subject of
controversy and debate (Saltelli, 2007). Equal
weighting of indicators implies that all variables are
worth the same in the composite indicator. However,
the decision to apply equal weighting could be the
consequence of the absence of a statistical or an
empirical basis or a lack of consensus on the
alternative (Giovannini et al., 2008). Paruolo et al.
(2013) noted that in many cases the declared
importance of single indicators and their main effect
are very different, and that the data correlation
structure often prevents developers from obtaining
the stated importance.
In order to reduce subjectivity, improve stability and
overall quality of the SSI ranking model, we apply I-
distance methodology on the SSI-2014 values (SSI,
2014). One task is to provide clear understanding of
how each indicator contributes to the final position
of different countries, as well as to identify the most
influential indicators and assess their relative influ-
ence on the overall results. The second goal is to use
analysis of the relative contribution of the indicators
as a tool to examine possibilities for reduction of a
number of indicators in iterative assessment process
focused on gradual improvement of the statistical
quality of the model. By this, we provide in-depth
analysis and more comprehensive understanding of
specific factors that determines one country ranking
position. It could represent useful signals for
policymakers in which priority areas they should
focus relevant measures and programs to have the
most efficient impact on the overall positioning of
the country.
The paper is structured as follows. Section 2 focuses
on the post hoc I-distance methodology. The results
of the analysis are discussed in Section 3. The last
section provides a summary of the conclusions, in-
cluding policy implications.
2. Methodology
The often problem with a different composite indi-
cator based ranking and rating methods is that nor-
mative and subjective model assumptions signifi-
cantly affect the results of measurements. This
problem can be overcome by the use of the statistical
I-distance method, which was originally developed
by Ivanovic (1973, 1977) and which has been
recently significantly advanced (Jeremic et al., 2012;
Maricic et al, 2014; Dobrota et al., 2015a,b;
Isljamovic et al., 2015). The I-distance measurement
is based on calculating the mutual distances between
the entities, where I-distance is a metric distance in
an n-dimensional space. Any entity with real or fic-
tive minimal, maximal or average values of all its
variables’ values can be considered as the referent
entity. The ranking of assessed entities in the set is
based on the calculated distance from the referent en-
tity. For a designated set of variables XT = (X1, X2,
. . . Xk) that characterize the entities under assess
Electronic copy available at: https://ssrn.com/abstract=2856356
127
ment, the I-distance between the two entities er =
(x1r, x2r, . . . xkr) and es = (x1s, x2s, . . . xks) is given
as
d i r , s  i 1
 1  r 
k
D(r , s)   ji.12... j 1
i 1  i j 1
where di(r,s) represent the distance between the
values of variable Xi for er and es, e.g. the
discriminate effect,
di(r,s) = xir – xis , i1, ... , k
i the standard deviation of Xia and rji.12...j-1 is a
partial coefficient of the correlation between Xi and
Xj, (ji).
The measure of the square I-distance represents a
solution to the problem of the negative coefficient of
partial correlation, which can occur in cases where is
not possible to achieve the same direction of all
variables in all sets. The square I-distance is defined
as:
d 2 r , s  i 1
 
k
D 2 r , s    i 2  1  rji2.12... j 1
i 1 i j 1
Based on CIDI approach (Dobrota et al., 2015a), I-
distance could be used to determine indicator
weights based on the empirical Pearson correlations,
where values of correlations are divided by the sum
of correlations:
ri
wi  j
 rj
r 1
where ri (i = 1, . . . k) is a Pearson correlation
between i-th input variable and I-distance value and
the final sum of all weights equals 1 (same applies
using coefficient of determination, as used in our
case study).
3. Results
Our analysis has been performed in several consec-
utive I-distance iterations, starting with the first I-
distance iteration that includes all of the indicators
listed in Table 1, together with calculated coeffi-
cients of determination. For the following I-distance
iterations, a variable with the smallest coefficient of
determination in the previous iteration has been ex-
cluded from the further analysis. The analysis of the
relative contribution of the indicators to the score of
some country has been used to determine the final
number of iterations. This statistical method could
provide information if some indicators dominate
over the total scores and what is the level of that in-
fluence (Saisana and d’Hombres, 2008). The relative
contributions have been calculated as a proportion of
an indicator score multiplied by the appropriate
weight (calculated by the CIDI approach) with
regard to the overall score for each indicator and
country. Average relative contributions and standard
deviations have been calculated in next step, where
is important to note that higher standard deviation
leads to higher level of rank oscillation. Focus is to
128 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134

A
Table 2. Average relative contributions (Me) and related standard deviations (SD) for all variables (Var) comprehended in 14th
and 15th iterations
Var var19 var5 var9 var4 var13 var15 var3 var2 Total
14th iteration Me 0.1432 0.1201 0.1428 0.1481 0.0949 0.1212 0.1162 0.1135 1.0
SD 0.0158 0.0442 0.0189 0.0186 0.0269 0.0664 0.0701 0.0619 0.3228
Var var19 var9 var5 var4 var13 var15 var3 Total
15th iteration Me 0.1295 0.1674 0.1715 0.1156 0.1457 0.1358 0.1344 1.0
SD 0.0487 0.0232 0.0226 0.0329 0.0787 0.0808 0.0741 0.3609

Table 3. List of comprehended variables with their coefficient of determination (r2) and average r2 for initial and final iterations
1st iteration r2 2nd iteration r2 … 13th iteration r2 14th iteration r2
var9 0.7310 var9 0.7396 … var19 0.8911 var19 0.9025
var19 0.6115 var19 0.6448 … var5 0.8082 var5 0.8154
var5 0.5700 var5 0.5929 … var4 0.7039 var9 0.7709
var4 0.5155 var4 0.5655 … var9 0.7022 var4 0.7174
var13 0.4942 var13 0.5027 … var3 0.6889 var13 0.6708
var6 0.4638 var17 0.4998 … var15 0.6773 var15 0.6529
var17 0.4597 var6 0.4651 … var13 0.6773 var3 0.6480
var2 0.4290 var2 0.4529 … var2 0.6529 var2 0.6209
var15 0.4134 var15 0.4316 … var1 0.6384
… … … … …
var11 0.0317 var11 0.0259
var20 0.0066
Average r2 0.3316 0.3665 0.7156 0.7248

Table 4. Changes in rank, median rank and IQR for selected countries
Iterations
Country Delta Me IQR
1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th 11th 12th 13th 14th
Norway 1 1 1 1 1 2 2 2 2 3 2 6 1 1 0 2 1
Australia 21 24 24 29 27 24 12 13 13 13 12 8 2 2 -19 13 19
Netherlands 12 15 15 11 11 9 10 10 10 10 9 10 3 3 -9 10 7
Denmark 10 10 9 9 10 11 5 5 5 5 5 5 4 4 -6 5 4
Finland 4 4 4 3 4 4 1 1 1 1 1 3 5 5 1 4 3
… … … … … … … … … … … … … … … … … …
Lebanon 137 132 129 123 116 115 100 82 82 76 77 52 51 53 -84 91 57
… … … … … … … … … … … … … … … … … …
Sudan 151 151 151 151 151 151 150 150 150 149 149 149 149 147 -4 150 6
Congo DR 147 148 149 149 149 148 149 149 149 151 151 151 151 148 1 149 3
Niger 100 103 110 114 112 121 131 144 145 136 135 131 130 149 49 131 26
Sierra Leone 126 136 137 138 140 143 136 145 146 145 145 143 143 150 24 143 11
Chad 148 147 148 150 150 150 151 151 151 150 150 150 150 151 3 150 0

identify the iteration with the lowest cumulative crease fluctuations of the countries ranks, which
standard deviation of the average relative contribu- would decrease the quality of the model.
tions of each variable. This iteration should be the In our case, calculated values of cumulative standard
final one because following iterations would in- deviation have been continuously lower compared to

Electronic copy available at: https://ssrn.com/abstract=2856356

 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
the previous one for all iterations until the 15 th
iteration, which have had higher value then in 14th
iteration (Table 2). Therefore, the 14th iteration,
which comprehends eight statistically most im-
portant indicators, has been defined as the final one.
The in-depth analysis focused on the consistency in
the rank and general structure of the performance can
reveal the sensitivity of the rank of countries in
relation to each indicator included in the analysis.
Table 3 provides a review of comprehended varia-
bles, their coefficient of determination and appropri-
ate average coefficient of determination calculated
for initial and final iterations. The table with full re-
sults for all iterations is available with authors on re-
quest.
The average coefficient of determination continu-
ously rises through all iterations, confirming the
increase of the quality of the model (Markovic et al.,
2015). The results of the final 14th iteration are
pointing out the significant importance of eight
indicators out of initially considered 21 indicators,
which represent essential elements for utilization of
the I-distance framework in an evaluation of SSI
results and methodology. These most important
variables are: Gross Domestic Product (var19),
Healthy Life (var5), Good Governance (var9),
Education (var4), Energy Use (var13), Greenhouse
Gases (var15), Safe Sanitation (var3) and Sufficient
to Drink (var2). It can be noted that indicators from
all three dimensions are present in the final set, with
the highest share of indicators of human wellbeing
(5 indicators) compared to environmental (2
indicators) and economic wellbeing (1 indicator).
First ranked Gross Domestic Product is calculated
on GDP per capita PPP basis. GDP has been
qualified as a poor indicator of economic well-being
(Pissourios, 2013) which cannot be used alone to
describe the broader quality of life (Caminada et al.,
2010). However, the results of our analysis confirm
that GDP remains a very important indicator for
measuring the economic performance of countries,
which is a fundamental driver of well-being (Borini,
2012).
Health-adjusted life expectancy (HALE) at birth
represent a basis for the calculation of Healthy Life
variable. It represents a number of years that a
newborn is expected to live, reduced by the number
of years spent in poor health. WHO introduced this
metric with the goal to take into account not only the
average life expectancy of people but also their
health (Mathers et al., 2010).
Good Governance variable represents the average of
values of the six World Bank’s Governance
Indicators: Voice and Accountability, Political
Stability, Government Effectiveness, Regulatory
Quality, Rule of Law and Control of Corruption
(WGI, 2015). It is founded on the rationale that good
governance represents condition for the development
of all people in freedom and harmony, within the
Electronic copy available at: https://ssrn.com/abstract=2856356
129
framework of rules and laws (van de Kerk and
Manuel, 2014).
Education indicator is calculated based on the
UNESCO metric of combined gross enrollment ratio
(UNESCO, 2015). This ratio expresses the count of
students enrolled in primary, secondary and tertiary
levels of education, irrespective of their age and
expressed as a percentage of the population of
official school age for the all three levels. Outputs of
our study are in alignment with the opinions that
education plays crucial role in adaptation towards
sustainable development (Svanstrom et al., 2008)
and that it is essential for the change in social
attitudes that will be needed to protect the welfare of
future generations (Chalkley et al., 2013).
“Energy Use is calculated based on IEA measure of
energy consumption in tons oil equivalents (toe) per
person, where energy use higher than 5 toe per
capita is penalized with a zero value of the Energy
Use indicator (SSI, 2014). Energy is not only a key
requirement for economic progress, but also it causes
significant pressures on the environment, both by
depletion of the resources and by the creation of the
pollution. Results of our analysis confirm relevance
and importance of the ability of this variable to
reflect the energy-use patterns and aggregate energy
intensity of a society (IAEA, 2005).
Calculation formula for Greenhouse Gases as a
measure of main human contribution to climate
change is based on IEA metric of CO2 emissions per
capita per year, with zero-score penalization of the
emissions higher than 10 (SSI, 2014). As the latest
carbon dioxide emissions continue to track the top
end of emission scenarios (Piters et al., 2013), the
main focus in climate change battle is moving
towards economy-wide emission reduction targets
(Olivier et al., 2013).
Safe Sanitation variable is related to WHO
estimation of the share of total population with
sustainable access to improved sanitation,
considering connection to facilities like public
sewer, septic tank, pour-flush latrine, etc. (WHO,
2015). Accessibility to adequate sanitation facilities
is fundamental to decrease the risk and the frequency
of associated diseases.
Sufficient to Drink variable is based on another
WHO measure of the number of people as share of
the total population with sustainable access to an
improved water source like household connections,
public standpipes, protected wells, springs,
rainwater collection, etc. (WHO, 2015). The ability
of the latest two variables to describe general
hygiene and quality of life and their easy association
with other socioeconomic characteristics (like
education and income) make them useful indicators
of human development (UNDESA, 2007).
Table 4 provides changes in ranks of countries
between the first and the final iteration (Delta),
together with related median ranking position (Me)
130 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
and interquartile range (IQR) values, for the selected
countries. The full list of values for all countries is
available with authors on request. The interquartile
range is a robust measure of statistical dispersion that
represents the middle 50 percent of the distribution
of an ordered range of data. Being equal to the
difference between the upper and lower quartiles,
IQR is not affected by extreme values. Median pre-
sents the average value that falls in a middle of the
set of an ordered range of data.
In total, there are nine countries with a change of
rank for 50 or more places, which implies that these
countries are the most sensitive to excluding varia-
bles through iterations. The largest change in rank
occurs with Lebanon for 84 places. On the other
hand, small changes of the rank (max. 5 places) are
noted for 40 countries, including four of them (Nor-
way, New Zealand, Kenya, Guinea) without any
change.
Figure 1 shows the oscillation in the ranks of the five
top-ranked countries. Obviously, all of them are de-
mocracies with the highly developed economies.
Interestingly, Norway does not change its first
position between initial and last iteration, which can
be explained with very good scores in terms of the
elaborated most significant indicators: Safe
Sanitation, Sufficient to Drink (all 10/10), GDP
(9.89/10), Education (9.81/10), Healthy life
(8.99/10) and Good Governance (8.56/10).
Figure 1. Overview of the fluctuations in the ranks of the
five top-ranked countries
Figure 2. Overview of the fluctuations in the ranks of the
five bottom-ranked countries
Indeed, being one of the most developed countries
with very stable and prosperous mixed economy,
with both large state sector and a vibrant private
sector (NEP, 2014), Norway has very high value of
GDP per capita PPP of $65.640 (WB, 2015). In
addition to the best performance in area of sanitation
Electronic copy available at: https://ssrn.com/abstract=2856356
and drinking water, Norway have impressive
combined gross enrolment ratio of 98.1% and life
expectancy at birth equal to 74 years of healthy life
years (SSI, 2014). Although health care expenditure
participates in GDP with 9.4%, due to Norway’s
very high value of GDP per capita, its average health
expenditure is among the highest on a global level
(Ringard et al., 2013). Finally, the traditions and
institutions by which authority in a Norway is
exercised are perceived as among the best on global
level, which is illustrated by very high values of the
six Worldwide Governance Indicators (WGI, 2015).
At the same time, Norway has good performances in
many other variables that have been gradually
excluded, which explains the low level of rank oscil-
lations through iterations indicated by IQR value
equal to 1.
Dispersion of the rank measured by IQR for the most
of these top-ranked countries is reasonably low, ex-
cept for Australia (IQR=19). Australia is the only
one with a significant change of the rank by improv-
ing its position for 19 places, mainly in the 7th itera-
tion, where variable Biodiversity has been excluded.
There are almost 3000 threatened ecosystems and
ecological communities in Australia while 94% of
the bioregions have at least one or more threatened
ecosystems (Randall, 2008). At the same time, the
imposingly negative trend in forest area noticed for
an extended period (Bradshaw, 2012). All together it
could explain Australia’s poor SSI score (4.9) and a
very low 112th ranking position for Biodiversity
variable. On the other hand, Australia has quite good
scores in terms of Education, Safe Sanitation, Su-
fficient to Drink (all 10), GDP (9.61), Healthy Life
(9.12) and Good Governance (8.21). It implies that
good performances of Australia in previously elabo-
rated most significant indicators have been shado-
wed by the influence of less significant indicators
where Australia do not have impressing results.
Thus, reduction of the number of indicators provided
more transparent insight and enabled Australia to
improve its rank significantly through proposed iter-
ative procedure to the remarkable second overall po-
sition.
The fluctuations in the ranks of the five bottom-
ranked countries are illustrated in Figure 2. Chad is
a country in the last 151st position with a very small
change in rank (3 places) compared to the initial
iteration, mainly due to its quite poor scores for the
majority of the most significant indicators: Safe
Sanitation (1.19), GDP (1.73), Good Governance
(2.44) and Healthy Life (3.65).
Indeed, Chad is a poor country with weak
institutional and policy capacity, limited progress in
poverty reduction, fragile economy facing a trend
decline in oil revenues and vulnerable to oil price
and regional security shocks (IMF, 2014). Chad was
ranked in 184th place out of 187 countries on the
UN‘s 2014 HDI list (UNDP, 2014). Being one of the
poorest and most corrupted countries, it is noted that
 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
Chad failed to utilize recently emerged oil export
revenues to boost development and reduce poverty
due to three major factors: institutional capacity
constraints, socio-political incompatibilities and
subversive interactions with external factors
(Kojucharov, 2007).
On the other hand closely positioned Niger has no-
ticed one of the biggest increases in rank, which has
significantly worsened its position through iterations
for 49 places. With the negative shift of 19 places
and high level of dispersion of the ranking results
confirmed by IQR value equal to 26, the last iteration
appeared to be the most influential one, in which
variable Sufficient Food has been excluded from the
analysis. Obviously, Niger has good results in this
variable (9.36), and when its effect has been
neutralized, the final rank has been swiftly adjusted
in accordance with generally poor achievement in
the most of remaining variables. Niger is least-
developed, low-income country with desert counting
for two-thirds of its land and dominantly agricultural
population that is doubling every 18 years, which all
together represent a real challenge for food security,
healthcare, family planning, social protection,
education and employment growth (AEO, 2015).
The biggest change of rank for 84 places occurs with
Lebanon, causing improvement of its rank from
137th to 53rd position, and with the highest level of
dispersion of the result (IQR=57) measured in our
research. Its economy was hard-hit by the 15-years
civil war (1975–1990) and still occasionally suffers
from economic downturns caused by local and
regional political instabilities. At the same time, the
Lebanese real GDP has grown at a faster pace than
the regional real GDP since 2007 (Dagher and
Yacoubian, 2012). It can be noted that effects of
generally solid performances in the final set of
significant variables, which all have been better or
close to average values, have been extenuated with
relatively lower results in many of the other
variables. Among them following ones have had the
biggest negative impact: Biodiversity (score 3.85
compared to average per country 6.18; excluded in
7th iteration), Genuine Savings (score 3.66 compared
to average per country 6.57; excluded in 8th iteration)
and Gender Equality score 6.03 compared to average
per country 6.8; excluded in 12th iteration).
4. Conclusion
The rising popularity and importance of the sustain-
able development concept induced development of
numerous measurement approach and methodolo-
gies in the related area. Among them, the Sustainable
Society Index is currently the only one that attempts
to comprehend all three dimensions of sustainable
development integrally. This composite indicator is
constructed to be able to measure progress along all
three dimensions of sustainable development: hu-
man, environmental and economic.
Electronic copy available at: https://ssrn.com/abstract=2856356
131
In this paper, we explore possibilities to evaluate and
enhance SSI ranking calculation methodology based
on the use of an iterative multivariate post hoc I-dis-
tance approach. The I-distance method can synthe-
size many different indicators into a single numerical
value that represents the basis for ranking observed
entities. The proposed approach could provide not
only more comprehensive and detailed under-
standing of differences between countries but also a
deeper insight into the relative statistical importance
of selected indicators. This post hoc approach can be
a solid ground not only for gradual exclusion of the
less significant existing indicators but also for
inclusion of additional new variables in a process of
evaluation and improvement of the considered
measurement framework. The main idea behind pro-
posed approach is to improve the stability of the
ranking results and overall quality of the model, fo-
cusing on the analysis of the relative contribution of
the indicators in an iterative assessment process. By
this, we provide in-depth analysis and more compre-
hensive understanding of specific factors that deter-
mines one country ranking position.
As sustainability metrics are used as quantitative
information base and foundation for creation of
appropriate sustainable development policies, this
proposal has important policy implications related to
improvement of objectiveness of sustainability
metrics used in processes of sustainable develop-
ment policy making on all levels. Improved reliabil-
ity and accuracy of the sustainability metrics com-
bined with an analytical framework for exhaustive
understanding of the most significant factors influ-
encing one’s country positioning is of the highest im-
portance for sustainability policy making processes.
It is noted that statistical analysis of composite
indicators is essential to prevent media and
stakeholders taking them at face value possibly
leading to questionable policy choices (Paruolo et
al., 2013). More transparent and robust underpinning
statistical framework of composite indicators would
reduce perceived disproportion between strong com-
municative power and reliability of sustainability
indexes (Luzzati and Gucciardi, 2015). It would
simplify communication tasks of policymakers and
make sustainable development policies more
credible.
The proposed post hoc I-distance approach is able to
reveal specific priority areas for each country so that
policymakers could focus their attention on the areas
where relevant policies, programs and action plans
would have the most significant impact on the
overall relative position of the country. It could be
used to assess sustainable development in one coun-
try by the specific statistically-determined im-
portance level. Clear identification and focus on
those priority areas can represent a solid background
in planning and regular monitoring of progress to-
wards the strategic objectives of sustainable devel-
opment. It can provide significant inputs for policy
132 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
adjustment processes to identify and propose the
most efficient programs and measures. Policymakers
could use it to benchmark policies and to evaluate
consequence and influence of the alternative scenar-
ios on the country’s position relative to others.
There are many factors that explain reasons for the
current situation in one country, but obviously their
importance and significance cannot be equal. In our
approach, I-distance method is used to assess the rel-
ative contribution of the indicators in an iterative
process, resulting in eight SSI indicators that are
identified as the most significant. First ranked coun-
try is Norway, similar to its first position on the UN‘s
2014 HDI list (UNDP, 2014). It has superb perfor-
mances in the most of variables and dimensions,
coupled with the very low level of rank oscillations
through iterations. Although Norway does not have
the best scores noted for all variables, clearly it has a
better balance of the performances over a set of those
most crucial indicators compared to all other coun-
tries.
Obviously, it is important for policymakers to know
and understand their country’s relative position in
terms of the achieved level of sustainable develop-
ment per dimensions and overall. However, even
more important for them is to know in which areas
improvements are possible, how these improve-
ments would impact their positioning and in which
areas those improvements would provide the most
efficient effects on the sustainability ranking. The
proposed approach can improve capabilities of poli-
cymakers to understand the implications of selected
sustainability-related policies and their impacts on
the shaping and making development sustainable. As
in reality trade-offs are inevitable, there is an
imminent need for well-based and balanced choices
to be made during policy creation and adjustment,
where this approach could provide important insight
of relative importance and contribution of alternative
actions and measures in different areas.
Further potential study could be directed towards
assessment of each SSI dimension separately based
on CIDI approach, in order to better understand
relative importance of variables within correspond-
ing dimension and to exceed limitations of the
originally applied equal weighting scheme. Moreo-
ver, our approach could be the foundation for the
development of a more general framework for eval-
uation and improvement of some other composite in-
dicators concepts and approaches, not necessarily in
the area of sustainable development.
The main contribution of the paper is the application
of the I-distance methodology in an iterative process
focused on assessment of the relative contribution of
individual indicators to the final position of the
countries and appropriate stepwise reduction of
number of indicators. By this, we not only improve
stability and overall quality of the model but also
provide in-depth analysis and more comprehensive
understanding of specific factors that determines one
Electronic copy available at: https://ssrn.com/abstract=2856356
country ranking position. This is of crucial
importance policymakers to identify key indicators
and focus the priority areas where appropriate
policies, programs and measures would have the
most efficient impact on the overall positioning of
the country in terms of sustainable development.
References
1. AEO, Country note Niger, http://www.african
economicoutlook.org/en/country-notes/west-
africa/niger/ (4.08.2015).
2. BORINI R., 2012, Well-being and GDP: Why
we need them both. http://www.oecdbetterlife
index.org/blog/well-being-and-gdp.htm
(30.7.2015)
3. BRADSHAW C., 2012, Little left to lose:
deforestation and forest degradation in Australia
since European colonization, in: Journal of
Plant Ecology, vol. 5, no 1, p. 109-120.
4. CAMINADA K., GOUDSWAARD K., VAN
VLIET O., 2010, Patterns of welfare state
indicators in the EU: Is there convergence?, in:
Journal of Common Market Studies, vol. 48, no
3, p. 529-556.
5. CHALKLEY B., HAIGH M., HIGGITT D.,
2013, Education for Sustainable Development,
Routledge, London.
6. COBB C., HALSTEAD T., ROWE J., 1995,
The Genuine Progress Indicator: Summary of
data and methodology, Redefining Progress,
San Francisco, 1995.
7. DAGHER L., YACOUBIAN T., 2012, The
causal relationship between energy
consumption and economic growth in Lebanon,
in: Energy policy, vol. 50, p. 795-801.
8. DALY H. E., COBB J. B., 1989, For the
common good: redirecting the economy toward
community, the environment, and a sustainable
future, Beacon Press, Boston.
9. DOBROTA M., BULAJIC M., BORNMANN
L., JEREMIC V., 2015a, A new approach to the
QS university ranking using the composite I-
distance indicator: Uncertainty and sensitivity
analyses, in: Journal of the Association for
Information Science and Technology (in press).
10. DOBROTA M., MARTIC M., BULAJIC M.,
JEREMIC V., 2015b, Two-phased composite I-
distance indicator approach for evaluation of
countries’ information development, in:
Telecommunications Policy, vol. 39, no. 35, p.
406-420.
11. ESTY D., LEVY M., SREBOTNJAK T., DE
SHERBININ A., 2005, Environmental Sus-
tainability Index: Benchmarking National Envi-
ronmental Stewardship, Yale Center for Env-
ironmental Law & Policy, New Haven.
12. ESTY D., LEVY M., SREBOTNJAK T., DE
SHERBININ A., KIM C., ANDERSON B.,
2006, Pilot 2006 Environmental Performance
 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
Index, Yale Center for Environmental Law &
Policy, New Haven.
13. EWING B., MOORE D.; GOLDFINGER S.,
OURSLER A., REED A., WACKERNAGEL
M., 2010, The Ecological Footprint Atlas 2010,
Global Footprint Network, Oakland.
14. GIOVANNINI E., NARDO M., SAISANA M.,
SALTELLI A., TARANTOLA A., HOFFMAN
A., 2008, Handbook on constructing composite
indicators: methodology and user guide,
OECD, Paris.
15. HAMILTON K., ATKINSON G., PEARCE D.,
1997, Genuine Savings As An Indicator of
Sustainability, CSERGE working paper, UK
Economic and Social Research Council,
Swindon.
16. IAEA, 2005, Energy indicators for sustainable
development: guidelines and methodologies,
International Atomic Energy Agency, Vienna.
17. IMF, 2014, Chad Country Report No. 14/100,
http://www.imf.org/external/pubs/ft/scr/2014/
cr14100.pdf (3.08.2015).
18. ISLJAMOVIC S., JEREMIC V., PETROVIC
N., RADOJICIC Z., 2015, Colouring the socio-
economic development into green: I-distance
framework for countries welfare evaluation, in:
Quality & Quantity, vol. 49, 617-629.
19. IVANOVIC B., 1973, A Method of Establishing
a List of Development Indicators, UNESCO,
Paris.
20. IVANOVIC B., 1997, Classification Theory,
Institute for Industrial Economic, Belgrade.
21. JEREMIC V., BULAJIC M., MARTIC M.,
MARKOVIC A., SAVIC G., JEREMIC D.,
RADOJICIC Z., 2012, An evaluation of
European countries health systems through
distance based analysis, in: Hippokratia, vol.
16, no 2, p. 170-174.
22. KOJUCHAROV N., 2007, Poverty, Petroleum
& Policy Intervention: Lessons from the Chad-
Cameroon Pipeline, in: Review of African
Political Economy, vol. 34, p. 477-496.
23. LUZZATI T., GUCCIARDI G., 2015, A non-
simplistic approach to composite indicators and
rankings: an illustration by comparing the
sustainability of the EU Countries, in:
Ecological Economics, vol. 113, p. 25-38.
24. MARICIC M., JANKOVIC M., JEREMIC V.,
2014, Towards a Framework for Evaluating
Sustainable Society Index, in: Revista Română
de Statistică, vol. 3, p. 49-62.
25. MARKOVIC M., ZDRAVKOVIC M.,
MITROVIC M., RADOJICIC A., 2015, An
Iterative Multivariate Post Hoc I-Distance
Approach in Evaluating OECD Better Life
Index, in: Social Indicators Research (in press).
26. MATHERS C.D. et al., 2003, Methods for
Measuring Healthy Life Expectancy, in: Health
systems performance assessment: debates,
Electronic copy available at: https://ssrn.com/abstract=2856356
133
methods and empiricism, eds. Murray C.J.L.,
Evans, D., World Health Organization, Geneva.
27. MEBRATU D., 1998, Sustainability and
sustainable development: Historical and
conceptual review, in: Environmental Impact
Assessment Review, vol. 8, p. 493-520.
28. NEP, 2014, Norway Economy Profile 2014,
http://www.indexmundi.com/ norway/economy
_profile.html (1.08.2015).
29. OLIVIER J., JANSSENS-MAENHOUT G.,
PETERS J., 2013, Trends in global CO2
emissions – 2013 Report, Netherlands
Environmental Assessment Agency, The
Hague.
30. PARUOLO P., SAISANA M., SALTELLI A.,
2013, Ratings and rankings: Voodoo or
science?, in: Journal of the Royal Statistical
Society, Series A, vol. 176, no 3, p. 609-634.
31. PETROVIC N., 2012, Ecological management,
Faculty of Organizational Sciences, Belgrade.
32. PISSOURIOS I. A., 2013, An interdisciplinary
study on indicators: A comparative review of
quality-of-life, macroeconomic, environmental,
welfare and sustainability indicators, in:
Ecological Indicators, vol. 34, p. 420-427.
33. PITERS G., ANDREW R., BODEN T.,
CANADELL J., CIAIS P., LE QUERE C.,
MARLAND G., RAUPACH M., WILSON C.,
2013, The challenge to keep global warming
below 2°C, in: Nature Climate Change, vol. 3,
p. 4-6.
34. RANDALL A., 2008, Is Australia on a
sustainability path? Interpreting the clues, in:
Australian Journal of Agricultural & Resource
Economics, vol. 52, no 1, p.77-95.
35. RINGARD Å., SAGAN A., SPERRE SAUNES
I., LINDAHL A.K., 2013, Norway: health
system review, in: Health Systems in Transition,
vol. 15, no 8.
36. SAISANA M., D’HOMBRES B., 2008, Higher
education rankings: Robustness issues and
critical assessment. How much confidence can
we have in higher education rankings? Joint
Research Centre, Ispra.
37. SAISANA M., PHILIPPAS D., 2012,
Sustainable Society Index (SSI): Taking
societies’ pulse along social, environmental and
economic issues, Joint Research Centre, Ispra.
38. SALTELLI A., 2007, Composite Indicators
between analysis and advocacy, in: Social
Indicator Research, vol. 81, no 1, p. 65-77.
39. SALTELLI A., RATTO M., ANDRES T.,
CAMPOLONGO F., CARIBONI J., GATELLI
D., SAISANA M., TARANTOLA S., 2008,
Global sensitivity analysis, The Primer, John
Wiley & Sons, Chichester.
40. SSI, 2014, Sustainable Society Index – your
compass to sustainability, http://www.ssfindex.
com/ssi/ (1.7.2015).
134 Savić et al./Problemy Ekorozwoju/Problems of Sustainable Development 1/2016, 125-134
41. SVANSTROM M., LOZANO-GARCIA F.J.,
ROWE D., 2008, Learning outcomes for
sustainable development in higher education, in:
International Journal of Sustainability in
Higher Education, vol. 9, no 3, p. 339-351.
42. UN, 2005, World Summit Outcome, Resolution
A/60/1, http://data.unaids.org/ Topics/Universal
Access/worldsummitoutcome_resolution_24oc
t 2005_en.pdf (4.07.2015).
43. UNDESA, 2007, Indicators of Sustainable
Development: Guidelines and Methodologies,
Third edition, UN, New York.
44. UNDP, 2014, Human Development Report
2014 – Sustaining Human Progress: Reducing
Vulnerabilities and Building Resilience, United
Nations Development Programme, New York.
45. UNESCO, Education Indicators Technical
guidelines, http://www.uis.unesco.org/Library/
Documents/eiguide09-en.pdf (15.07.2015).
Electronic copy available at: https://ssrn.com/abstract=2856356
46. VAN DE KERK G., MANUEL A., 2008, A
comprehensive index for a sustainable society:
The SSI – the Sustainable Society Index,
http://www.sd-network.eu/pdf/resources/Sust%
20Society%20Index%20SSI%202008.pdf
(13.07.2015).
47. VAN DE KERK G., MANUEL A., 2014,
Sustainable Society Index 2014, Sustainable
Society Foundation, The Hague.
48. WB, World Databank Portal, http://databank.
worldbank.org/data/home.aspx (29.07.2015)
49. WCED, 1987, Our Common Future, Oxford
University Press, Oxford, 1987.
50. WGI, Worldwide Governance Indicators portal,
http://info.worldbank.org/governance/wgi/inde
x.aspx#doc (15.07.2015)
51. WHO, Definitions of Indicators, http://www.
who.int/water_sanitation_health/monitoring/
jmp04_2.pdf (22.07.2015).